"standard" pyro and deforming mesh

[Wolrajh]

Hey there,

I have some panicked people to set on fire. Most of the methods I've seen so far ( and used myself ) were 'hacks', meaning direct temperature field manipulation and/or advection of a custom field.

HOWEVER

I decided to give the standard combustion model AKA Pyrosolver a try in the hopes I would get more natural and physically correct burning. And well I'm running into quite a lot of issues, one of the most annoying being that the fuel field doesn't 'stick' to the animated / deforming mesh, resulting in the fuel staying up in the air like a trail until it's all burned out. Likewise, temperature and burn do not seem to follow, resulting in flames not sticking to the deforming mesh either, lagging behind since the leftover fuel ignites with a delay.

In short, it's a mess, and it's also super slow.

So basically I guess my question would be : is the Houdini Pyro combustion model a good fit for burning the hell out of screaming, kicking, little CG people, or is it just too complicated and slow and "locked" ? I've been getting correct results on props and set elements but it looks like a struggle for animated meshes. That's maybe why so many people would rather manipulate temperature ?

 

And if pyro is actually very capable in this situation, how can I make sure the fire correctly sticks to the mesh ? Add a few advection nodes ? Summon an old god ? Summon an old goat ?

[celd]

The easiest way to solve this is to set Burn Rate in combustion tab to 1, that way it wont keep any of the fuel for the next frame. There is also Advect Fuel toggle in Fuel tab but this probably wont be as helpful.

[Wolrajh]

Ah yes, indeed it works better. It's not entirely flawless, though - there is still that annoying lag behind as if the flames were pursuing the model limbs rather than being attached to it. Fun fact, I did try to crank up the Burn Rate a lot higher but I never tried settng the maximum, and the result is faaar different between 0.99 and 1.0 !

Anyway, in the end, after letting it rest for the night and trying to make it work for a few days, I'm quite dissatisfied with the provided pyro workflow. I'm quite sure I'm not entirely wrong about how I'm doing things but it requires so many patches and fixes to work 'well' that it ends up taking too much time simulating only to get a clunky result. I guess I'm better off with the 'fake' fire methods, temperature field manipulation and all that jazz. I mean, it's like building a house from scratch rather than trying to plug the leak in an old manor. Or something like that.

I may be wrong though, and I'd be curious to hear about people successfully getting nice and realistic fires using the pyrosolver model...